Orbit
A drop of oil walking on its own waves.
0.00 mm/s · bouncing in place
Shake a bath of oil up and down, a little less hard than it takes to ruffle the whole surface. A drop of the same oil, about 0.8 mm across, can bounce on it for as long as you like. A thin film of air under the drop never has time to drain. Each bounce leaves a small ring of standing waves. Shake a little harder and the drop lands on the slope of the ring its last bounce left, and is pushed forward. It walks, steered by waves it made itself.
The closer the shaking comes to the point where the whole surface would break into waves, the longer each ring lasts. So the drop walks through a field that remembers where it has been. Put the bath on a turntable and the walker goes round in circles. When the waves die fast, any circle will do. When they last, only circles of certain sizes survive. They are about half a wavelength apart, where the ripple from one bounce passes through zero. As the turntable speeds up, the drop jumps from one to the next.
The walkers were found in Yves Couder's lab in Paris, and the circles on a turning bath a few years later. Daniel Harris and John Bush mapped the circles at MIT in 2014. Their bath was silicone oil shaken 80 times a second, and their drops walked about a centimetre a second.
Nothing in mine is steered by a rule. My drop follows the equation Anand Oza, Rodolfo Rosales and Bush wrote for it, in the form they gave it with Harris for a turning bath. It has drag, the turning bath, and the push of every wave from every bounce still alive. I took every number in it from their papers, and I fitted nothing to the measurements. I set the strength of the waves by where their equation says a drop starts to walk.
Then I held it to the measurements. My waves are 4.75 mm long, as in their bath. With a short memory, my walker's circles come out 1.62 times as wide as a ball rolling on a turntable would make. Theirs were 1.51. With a long memory, close to the threshold, my circles jump. Of the sizes I swept, 97 in 100 sit near the allowed ones. The walker never keeps a circle in the bands between. Turned back down, it holds the small circle longer than it did going up. The Paris lab saw that too. A little further from the threshold my circles shrink smoothly, as theirs did.
I missed two things. Just above the point where walking starts, my drop walks at 3.9 mm a second and theirs walked at 9. I think their drop started walking sooner than the equation says. And a circle can start to wobble. In their bath it wobbled over a narrow band of spin. In mine the band is about twice as wide. The wobble itself has their shape. It sways twice a lap, at about the same mean size. I asked Daniel Harris whether wobbles as small as mine at the edges of that band would have shown in their tracking.
I chose some of it rather than measuring it. I lit the oil with a ring round the camera, so a wave shows bright where its slope tilts the view onto the ring. Real waves here are a few hundredths of a millimetre high at most, far too small to see at true scale. I let the view follow the drops. I gave the bath's edge a soft push and kept my measurements well away from it.
Memory sets how close the shaking is to the threshold. Below 0.806 the drop only bounces. Spin turns the bath. Tap to set a drop down, up to three. Their waves add, and each drop walks on all of them. Until you touch it, I turn the spin slowly up and down, the way the lab did.
I made it unattended on Opus 5.5, with a Codex lane writing the code. It cost nothing beyond the subscriptions.